Search Results

Search found 8224 results on 329 pages for 'sometimes'.

Page 284/329 | < Previous Page | 280 281 282 283 284 285 286 287 288 289 290 291  | Next Page >

• VS2010 + IE8 Debugging woes - Element not found

  - by Chin

  I am having great difficulty trying to debug with vs2010 and IE8, though I think the problem is more IE8 specific. When starting a debug session 9 times out of 10 I will have the following problem. IE tab says connecting.. - then after a 5 second wait I will get an error in VS saying element not found. Even when I click ok to dismiss the error, the IE window still shows connecting... I will then have to kill the IE process to be able to close IE to try again. Sometimes however I am lucky and it starts. But the whole thing is so random I have no clue where to start. One thing I have noticed is that I always have 2 IE processes started even though there is only one window open. One has a small footprint of 100k, I presume it is some kind of helper. I am using a static port with the built in WebDev server. If anyone has had similar problems please let me know how you resolved it. Its driving me nuts! thanks

  Read the article

• Frame sizing of tableview within nested child controllers/subviews

  - by jwoww

  I'm a bit confused by the proper frame sizing of a table view to fit within my screen. Here's my setup of view controllers within view controllers: UITabBarController UINavigationController as one of the tab bar viewcontrollers; title bar hidden ViewController - a container view controller because I need the option to place some controls beneath the UITableView, sometimes (but not in the current scenario) UITableViewController Now, my question is what the proper frame dimensions of the UITableview should be. Here's what I've got in the ViewController viewDidLoad method. I used subtracted 49.0 (the size of the tab bar) from 480.0. However, this leaves a black bar at the bottom. 20.0 appears to do it (coincidentally?) the size of the status bar, but I don't understand why that would be. Wouldn't the true pixel dimensions of the tableview be 480-49? // MessageTableViewController is my subclass of UITableViewController MessagesTableViewController *vcMessagesTable = [[MessagesTableViewController alloc] init]; CGRect tableViewFrame = CGRectMake(0, 0, 320.0, 480.0 - 49.0); [[vcMessagesTable view] setFrame:tableViewFrame]; self.tableViewController = vcMessagesTable; [self addChildViewController:vcMessagesTable]; [[self view] addSubview:vcMessagesTable.view]; Here's how it looks:

  Read the article

• Is "tip-of-the-day" good?

  - by Jonta

  Many programs (often large ones, like MS Office, The GIMP, Maxthon) have a feature called "tip-of-the-day". It explains a small part of the program, like this one in Maxthon: "You can hide/show the main menu bar by pressing Ctrl+F11" You can usually browse through them by clicking next. And other options provided are "Previous", "Close", "Do not show at startup". I think I like the way Maxthon used to handle this; in the browser's statusbar (down at the bottom usually, together with "Done", the progress-bar etc), there would sometimes be a small hint or tip on what else you could do with it. As Joel Spolsky wrote in his article-series "User Interface Design for Programmers", people don't like reading manuals. But we still want them to use the program, and the features they could benefit from, don't we? Therefore, I think it is useful to have such a feature, without the annoyance of the pop-up on startup. What do you think? Pop-up? Maxthonstyle? No way?

  Read the article

• Looking for an appropriate design pattern

  - by user1066015

  I have a game that tracks user stats after every match, such as how far they travelled, how many times they attacked, how far they fell, etc, and my current implementations looks somewhat as follows (simplified version): Class Player{ int id; public Player(){ int id = Math.random()*100000; PlayerData.players.put(id,new PlayerData()); } public void jump(){ //Logic to make the user jump //... //call the playerManager PlayerManager.jump(this); } public void attack(Player target){ //logic to attack the player //... //call the player manager PlayerManager.attack(this,target); } } Class PlayerData{ public static HashMap<int, PlayerData> players = new HashMap<int,PlayerData>(); int id; int timesJumped; int timesAttacked; } public void incrementJumped(){ timesJumped++; } public void incrementAttacked(){ timesAttacked++; } } Class PlayerManager{ public static void jump(Player player){ players.get(player.getId()).incrementJumped(); } public void incrementAttacked(Player player, Player target){ players.get(player.getId()).incrementAttacked(); } } So I have a PlayerData class which holds all of the statistics, and brings it out of the player class because it isn't part of the player logic. Then I have PlayerManager, which would be on the server, and that controls the interactions between players (a lot of the logic that does that is excluded so I could keep this simple). I put the calls to the PlayerData class in the Manager class because sometimes you have to do certain checks between players, for instance if the attack actually hits, then you increment "attackHits". The main problem (in my opinion, correct me if I'm wrong) is that this is not very extensible. I will have to touch the PlayerData class if I want to keep track of a new stat, by adding methods and fields, and then I have to potentially add more methods to my PlayerManager, so it isn't very modulized. If there is an improvement to this that you would recommend, I would be very appreciative. Thanks.

  Read the article

• Reversing strings in a vector using for_each and bind

  - by fmuecke

  Hi! I was wandering how it's possible to reverese strings that are contained in a vector using a single for_each command just in one "simple" line. Yea, I know it is easy with a custom functor, but I can't accept, that it can't be done using bind (at least I couldn't do it). #include <vector> #include <string> #include <algorithm> std::vector<std::string> v; v.push_back("abc"); v.push_back("12345"); std::for_each(v.begin(), v.end(), /*call std::reverse for each element*/); Edit: Thanks a lot for those funtastic solutions. However, the solution for me was not to use the tr1::bind that comes with the Visual Studio 2008 feature pack/SP1. I don't know why it does not work like expected but that's the way it is (even MS admits that it's buggy). Maybe some hotfixes will help. With boost::bind everything works like desired and is so easy (but sometimes relly messy:)). I really should have tried boost::bind in the first place...

  Read the article

• Operator overloading in generic struct: can I create overloads for specific kinds(?) of generic?

  - by Carson Myers

  I'm defining physical units in C#, using generic structs, and it was going okay until I got the error: One of the parameters of a binary operator must be the containing type when trying to overload the mathematical operators so that they convert between different units. So, I have something like this: public interface ScalarUnit { } public class Duration : ScalarUnit { } public struct Scalar<T> where T : ScalarUnit { public readonly double Value; public Scalar(double Value) { this.Value = Value; } public static implicit operator double(Scalar<T> Value) { return Value.Value; } } public interface VectorUnit { } public class Displacement : VectorUnit { } public class Velocity : VectorUnit { } public struct Vector<T> where T : VectorUnit { #... public static Vector<Velocity> operator /(Vector<Displacement> v1, Scalar<Duration> v2) { return new Vector<Velocity>(v1.Magnitude / v2, v1.Direction); } } There aren't any errors for the + and - operators, where I'm just working on a Vector<T>, but when I substitute a unit for T, suddenly it doesn't like it. Is there a way to make this work? I figured it would work, since Displacement implements the VectorUnit interface, and I have where T : VectorUnit in the struct header. Am I at least on the right track here? I'm new to C# so I have difficulty understanding what's going on sometimes.

  Read the article

• How to use git to manage one codebase but have different environments

  - by emostar

  I'm using git for a personal project at the moment and have run into a problem of having one codebase for two different environments and was wondering what the cleanest way to use git would be. Main Desktop I Use this machine for most of my development. I have a git repository here that I cloned off of an empty repository that I use on my internal server. I do most of my work here and push back to the internal server so I can use that as a master of truth and to ease making backups. Laptop I sometimes want to code on the road, so I did a clone from the internal server and created a new branch called "laptop-branch". Unfortunately some directories MSVC++ version are different than from the Main Desktop environment. I just modified the files in the "laptop-branch" and committed them there. Now I did a lot of changes while on vacation with my laptop, and want to push them to origin, but don't want the changes I made that were related to directories and compiler versions to be pushed back to origin. What would be the best way to get this done?

  Read the article

• jQuery: Problems cookies internet explorer

  - by user1140479

  I have made a login page. When the user logs in a request to an API is send. This API is PHP and checks the username and password. When both are correct an unique key is send back (this is placed in the database for further use: userid and other stuff needed in the website). After that key is sent back it is placed in a cookie: $.cookie("session", JSON.stringify(result)); After the cookie is set I send the user to a new page: location.href = 'dashboard.htm'; In this page jQuery checks if the cookie "session" is present. If not, the user is send back to the login page. sessionId = ($.cookie("session") ? JSON.parse($.cookie("session")).SessionId : 0); return sessionId; This works fine in Chrome, but IE (8/9) has some problems with this. I figured out that when you get to dashboard.htm the session is present. As soon as I hit F5 the session is gone. And sometimes the cookie isn't set at all! I can't seem to figure out why this is happening in IE. Has someone any idea? Other options/ideas to save that unique key are also welcome. Thanks in advance.

  Read the article

• Zend Framework - How do make a hierarchy without it being a module?

  - by Josh

  Here is my specific issue. I want to make an api level which then under that you can select which method you will use. For example: test.com/api/rest test.com/api/xmlprc Currently I have api mapping to a module directory. I then setup a route to make it a rest route. test.com/api is a rest route, but I would rather have it be test.com/api/rest. This would allow me later to add others. In Bootstrap.php: $front = Zend_Controller_Front::getInstance(); $router = $front->getRouter(); $route = new Zend_Controller_Router_Route( 'api/:module/:controller/:id/*', array('module' =>'default') ); $router-addRoute('api', $route); $restRoute = new Zend_Rest_Route($front, array(), array( 'rest' )); $router-addRoute('rest', $restRoute); return $router; In application.ini: resources.frontController.moduleDirectory = APPLICATION_PATH "/modules" I know it will involve routes, but sometimes I find the Zend Framework documentation to be a little hard to follow. When I go to test.com/rest/controller/ it works how it should, but if I go to test.com/api/rest/ it tells me that my module is api and controller is rest.

  Read the article

• Selenium - Wait for network traffic

  - by Joel

  We're using Selenium with the Java API and some Javascript user extensions. We use a lot of AJAX calls in our app. A lot of our tests fail randomly because sometimes the AJAX calls finish slower than other times so the page isn't fully loaded. We fix that by waiting for specific elements or Thread.sleep. I was trying to find a way to instead just wait for the network traffic to finish. So that we could do this: selenium.click("some JS button"); selenium.waitForNetwork(); assertTrue(something); That way we can get rid of the thread sleep and have tests pass faster when the server responds faster and not have so many tests fail due to timing issues. I haven't been able to find a way to do this searching Google. Does anyone have any ideas how we can accomplish this? (Preferably either through Javascript or the Java API but all suggestions are welcome). Note: the other variations of "waitFor" are not what I'm looking for. We're already using those in clicks and other things. I'm looking for something that waits for the NETWORK TRAFFIC. Thanks for all the feedback, I'll be trying out a couple of the suggestions, but I'm still open to other ideas. Thanks.

  Read the article

• Application HANGS UP if 3G connection is used (WiFi is OK)

  - by SkyEagle888

  Hi, My application uses the following codes to connect to web site and download a web page for processing. HttpURLConnection c = (HttpURLConnection) url.openConnection(); c.setRequestMethod("GET"); c.setRequestProperty("User-Agent", "Mozilla/3.0 (compatible; MSIE 4.0; Windows NT)"); c.setRequestProperty("Accept-Language", "zh-hk"); c.setRequestProperty("Connection", "Keep-Alive"); c.setConnectTimeout(6000); // mTimeout is in seconds intTries = 3; do { c.connect(); if (c.getResponseCode() == 200) { break; } else { intTries --; } } while (intTries > 0); The codes are run in main thread only. If WiFi is being used, it is nice and clean. If 3G connection is being used, sometimes, it cannot get connected OR application HANGS. Is there any significant difference between handling WiFi and 3G ? Thanks.

  Read the article

• Should the program logic reside inside the gui object class or be external to the class?

  - by hd112

  I have a question about how to structure code in relation to GUI objects. Suppose I have a dialog that has a list control that has a bunch of names obtained from a database. The user can edit the names. Does the logic reside inside that dialog class or should it be from the outside. To illustrate what I mean, here’s some pseudo code showing the structure of the code when the logic is handled outside the dialog class: NamesDialog : wxDialog { Private: ..stuff.. Public: ... SetNames(wxStringArray names); wxStringArray GetNames(); ..stuff.. } So the user of the class would do something like: wxStringArray names = DatabaseManager::Get()->GetNames(); names.Sort(); NamesDialogObject.SetNames(names); NamesDialogObject.ShowModal(); wxStringArray modified_names = NamesDialogObject.GetNames(); AddToDatabase(modified_names); //or something like this. On the other hand, the database logic can reside inside the NamesDialog class itself. In the show method I can query the database for the names and as the user interacts with the controls (list control in this case), the database can be updated from the event handlers. As a result the NamesDialog class only has the Show() method as there is no need to use SetNames or GetNames() etc. Which method is generally preferred? I don’t have much work experience so I’m not sure which is the proper way to handle it. Sometimes it's easier to handle everything in the class but getting access to objects it interacts with can be challenging. Generally can do it by having the relevant objects be singletons like the database manager in the above example.

  Read the article

• Split a string by comma, quote and full-stop.. with a few exceptions

  - by dunc

  I've got a lot of text, similar to the following paragraph, which I'd like to split into words without punctuation (', ", ,, ., newline etc).. with a few exceptions. Initially considered endemic to the Chalakudy River system in Kerala state, southern India, but now recognised to have a wider distribution in surrounding drainages including the Periyar, Manimala, and Pamba river though the Manimala data may be questionable given it seems to be the type locality of P. denisonii. In the Achankovil River basin it occurs sympatrically, and sometimes syntopically, with P. denisonii. Wild stocks may have dwindled by as much as 50% in the last 15 years or so with collection for the aquarium trade largely held responsible although habitats are also being degraded by pollution from agricultural and domestic sources, plus destructive fishing methods involving explosives or organic toxins. The text refers to P. denisonii which is a species of fish. It's an abbreviation of Genus species. I would like this reference to be one word. So, for instance, this is the kind of array I'd like to see: Array ( ... [44] given [45] it [46] seems [47] to [48] be [49] the [50] type [51] locality [52] of [53] P. denisonii [54] In [55] the ... ) The only things that distinguish these species references such as P. denisonii from a new sentence like end. New are: The P (for Puntius, as in the P. in the aforementioned example) is only ever one letter, always a capital the d (as in . denisonii) is always either a lower case letter or an apostrophe (') What regexp can I use with preg_split to give me such an array? I've tried a simple explode( " ", $array ) but it doesn't do the job at all. Thanks in advance,

  Read the article

• GTK+: How do I process RadioMenuItem choice without marking it chosen? And vise versa

  - by eugene.shatsky

  In my program, I've got a menu with a group of RadioMenuItem entries. Choosing one of them should trigger a function which can either succeed or fail. If it fails, this RadioMenuItem shouldn't be marked chosen (the previous one should persist). Besides, sometimes I want to set marked item without running the choice processing function. Here is my current code: # Update seat menu list def update_seat_menu(self, seats, selected_seat=None): seat_menu = self.builder.get_object('seat_menu') # Delete seat menu items for menu_item in seat_menu: # TODO: is it a good way? does remove() delete obsolete menu_item from memory? if menu_item.__class__.__name__ == 'RadioMenuItem': seat_menu.remove(menu_item) # Fill menu with new items group = [] for seat in seats: menu_item = Gtk.RadioMenuItem.new_with_label(group, str(seat[0])) group = menu_item.get_group() seat_menu.append(menu_item) if str(seat[0]) == selected_seat: menu_item.activate() menu_item.connect("activate", self.choose_seat, str(seat[0])) menu_item.show() # Process item choice def choose_seat(self, entry, seat_name): # Looks like this is called when item is deselected, too; must check if active if entry.get_active(): # This can either succeed or fail self.logind.AttachDevice(seat_name, '/sys'+self.device_syspath, True) Chosen RadioMenuItem gets marked irrespective of the choose_seat() execution result; and the only way to set marked item without triggering choose_seat() is to re-run update_seat_menu() with selected_seat argument, which is an overkill. I tried to connect choose_seat() with 'button-release-event' instead of 'activate' and call entry.activate() in choose_seat() if AttachDevice() succeeds, but this resulted in whole X desktop lockup until AttachDevice() timed out, and chosen item still got marked.

  Read the article

• JQuery event that triggers after CSS is loaded?

  - by Dave

  (I posted this on the jquery forums but it's still awaiting moderation, so I thought I'd try my luck here since stackoverflow is so awesome. If I get an answer I'll post it here.) Maybe someone can help me with this, I have a couple of links on my page (inside a <div id="theme-selector">) which allow you to change the CSS stylesheets: $('#theme-selector a').click(function(){ var path = $(this).attr('href'); $('head link').remove(); $('head').append('<link type="text/css" href="'+path+'" rel="stylesheet" />'); return false; }); Now, after I've changed the style on the page, I want to get the new background color, using the following code (which I put after the $('head').append call): var bgcolor = $('body').css('background-color'); alert(bgcolor); The problem is, I think, that it takes some time for the browser to download the new stylesheet and I sometimes get the old background color in my alert message. Is there some event I can bind that will only alert me after all the stylesheets are loaded on the page? At the moment, all I can think of is using a setTimeout(function(){}, 5000); which isn't great, because what if it takes longer/shorter to load all the CSS on the page. Let me know if I need to clarify anything and I can provide more code. Thanks in advance.

  Read the article

• Common protected data member in base class?

  - by EXP0

  I have a base class and several derived classes. The derived classes use some common data, can I just put those common data as protected member of the base class? I know the protected member breaks encapsulation sometimes, so I wonder if there is any good approach. Here is a specific example: class Base{ public: virtual void foo() = 0; void printData(); protected: std::vector<std::string> mData; } class Dr1 : public Base{ public: virtual void foo(); //could change mData } class Dr2 : public Base{ public: virtual void foo(); //could change mData } If I put mData into Dr1 and Dr2 as private member, then I need to put it in both of them, and I can not have printData() in Base since printData() need access to mData unless I make printData() virtual and have identical function in both Dr1 and Dr2, which doesn't make much sense to me. Is there a better way to approach this without using protected member? Thank you.

  Read the article

• understanding and implementing Boids

  - by alphablender

  So I'm working on porting Boids to Brightscript, based on the pseudocode here: http://www.kfish.org/boids/pseudocode.html I'm trying to understand the data structures involved, for example is Velocity a single value, or is it a 3D value, ie velocity={x,y,z} It seems as if the pseudocode seems to mix this up where sometimes it has an equation that incudes both vectors and single-value items: v1 = rule1(b) v2 = rule2(b) v3 = rule3(b) b.velocity = b.velocity + v1 + v2 + v3 If Velocity is a tripartite value then this would make sense, but I'm not sure. So my first question here is, is this the correct datastructure for a single boid based on the Pseudocode on that page: boid={position:{px:0,py:0,pz:0},velocity:{x:0,y:0,z:0},vector:{x:0,y:0,z:0},pc:{x:0,y:0,z:0},pv:{x:0,y:0,z:0}) where pc=perceived center, pv= perceived velocity I"ve implemented a vector_add, vector_sub, vector_div, and vector boolean functions. The reason I'm starting from this pseudocode is I've not been able to find anything else that is as readable, but it still leaves me with lots of questions as the data structures are not explicitly defined for each variable. (edit) here's a good example of what i'm talking about: IF |b.position - bJ.position| < 100 THEN if b.position - b[j].position are both 3D coordinates, how can they be considered "less than 100" unless they are < {100,100,100} ? Maybe that is what I need to do here, use a vector comparison function?

  Read the article

• What access token should i use?

  - by user548458

  I made a posting scores. It worked normaly sometimes. But, now I cant post any score. If I use user accessToken: string path = meID + "/scores"; var parameters = new Dictionary<string, object> (){ { "score", score.ToString () }, { "access_token", accessToken } }; facebook.post (path, parameters, ( error, obj ) => I get error: {"error":{"message":"(#15) This method must be called with an app access_token.", "type":"OAuthException", "code":15}} If I use an app access token: string path = meID + "/scores"; var parameters = new Dictionary<string, object> () { { "score", score.ToString () }, { "access_token", appAccessToken } }; facebook.post (path, parameters, ( error, obj ) => I get other error: {"error":{"message":"A user access token is required to request this resource.", "type":"OAuthException", "code":102}} Help me please, what am i doing wrong? PS: I worked well recently, but now - dont. Cant explain it...

  Read the article

• How do you get the solution directory in C# (VS 2008) in code?

  - by IsaacB

  Hi, Got an annoying problem here. I've got an NHibernate/Forms application I'm working through SVN. I made some of my own controls, but when I drag and drop those (or view some form editors where I have already dragged and dropped) onto some of my other controls, Visual studio decides it needs to execute some of the code I wrote, including the part that looks for hibernate.cfg.xml. I have no idea why this is, but (sometimes!) when it executes the code during my form load or drag and drop it switches the current directory to C:\program files\vs 9.0\common7\ide, and then nhibernate throws an exception that it can't find hibernate.cfg.xml, because I'm searching for that in a relative path. Now, I don't want to hard code the location of hibernate.cfg.xml, or just copy hibernate.cfg.xml to the ide directory (which will work). I want a solution that gets the solutions directory while the current directory is common7\ide. Something that will let someone view my forms in the designer on a fresh checkout to an arbitrary directory on an arbitrary machine. And no, I'm not about to load the controls in code. I have so many controls within controls that it is a nightmare to line everything up without it. I tried a pre build event that made a file that has the solution directory in it, but of course how can I find that from common7\ide? All the projects files need to be in the solution directory because of svn. Thanks for your help guys, I've already spent a few hours fiddling with this in vain.

  Read the article

• Socket Performance C++ Or C#

  - by modernzombie

  I have to write an application that is essentially a proxy server to handle all HTTP and HTTPS requests from our server (web browsing, etc). I know very little C++ and am very comfortable writing the application features in C#. I have experimented with the proxy from Mentalis (socket proxy) which seems to work fine for small webpages but if I go to large sites like tigerdirect.ca and browse through a couple of layers it is very slow and sometimes requests don't complete and I see broken images and javascript errors. This happens with all of our vendor sites and other content heavy sites. Mentalis uses HTTP 1.0 which I know is not as efficient but should a proxy be that slow? What is an acceptable amount of performance loss from using a proxy? Would HTTP 1.1 make a noticeable difference? Would a C++ proxy be much faster than one in C#? Is the Mentalis code just not efficient? Would I be able to use a premade C++ proxy and import the DLL to C# and still get good performance or would this project call for all C++? Sorry if these are obvious questions but I have not done network programming before.

  Read the article

• How to properly manage a complex DB structure?

  - by errr

  Let's say you have several systems using the same DB - each uses several schemes (sometimes same as the other). This structure of these schemes is somewhat very big and complicated. Now, how could you possibly manage such scheme structure? Obviously using some sort of "configuration" - the simplest would be SQL scripts, but a more reasonable solution would be XMLs which can be easily converted into SQL, or some other readable solution (for example, JPA's XMLs or Annotations). This solution though, causes a problem where you can't really tell if your configuration matches the structure of the DB schemes exactly. You can't say if those two are synchronized. Why wouldn't they? Well, in such big structure there are going to be many changes, and you won't always remember to save/commit your configuration after you've altered the schemes, or maybe you did save/commit it, but eventually didn't altered anything in the schemes and forgot to undo the changes to the configuration. More than that, another problem (not caused by the configuration, but isn't addressed by it either) is versioning. I don't see any good way of managing the DB schemes versions (say our last alteration makes 3 systems crash - not good, how to "rollback"?). And thoughts? thx.

  Read the article

• Problem displaying the Message box in MFC

  - by kiddo

  I have a simple MFC program which displays the progressbar..I used the below code to display the progress bar.. HWND dialogHandle = CreateWindowEx(0,WC_DIALOG,L"Proccessing...",WS_OVERLAPPEDWINDOW|WS_VISIBLE, 600,300,280,120,NULL,NULL,NULL,NULL); HWND progressBarHandle = CreateWindowEx(NULL,PROGRESS_CLASS,NULL,WS_CHILD|WS_VISIBLE|PBS_MARQUEE,40,20,200,20, dialogHandle,(HMENU)IDD_PROGRESS,NULL,NULL); while(FALSE == testResult) { MSG msg; SendMessage(progressBarHandle, PBM_SETRANGE, 0, MAKELPARAM( 0, 100 ) ); SendMessage(progressBarHandle,PBM_SETPOS,0,0); ShowWindow(progressBarHandle,SW_SHOW); Sleep(50); if(TRUE == myCondition)//myCondition is a bool variable which is decalred globally { DestroyWindow(dialogHandle); AfxMessageBox(L"Test Success"); } } when I execute the above code..the message box displays only after a mouseover event.like if I move the mouse the message box will display if not it will not display until i move the mouse. And also while the progressbar is running if I try to move the progress bar window..it displays a windows background at the place of displacement and also in the new region or sometimes its getting stuck.Please help me with this!

  Read the article

• Another thread safe queue implementation

  - by jensph

  I have a class, Queue, that I tried to make thread safe. It has these three member variables: std::queue<T> m_queue; pthread_mutex_t m_mutex; pthread_cond_t m_condition; and a push and pop implemented as: template<class T> void Queue<T>::push(T value) { pthread_mutex_lock( &m_mutex ); m_queue.push(value); if( !m_queue.empty() ) { pthread_cond_signal( &m_condition ); } pthread_mutex_unlock( &m_mutex ); } template<class T> bool Queue<T>::pop(T& value, bool block) { bool rtn = false; pthread_mutex_lock( &m_mutex ); if( block ) { while( m_queue.empty() ) { pthread_cond_wait( &m_condition, &m_mutex ); } } if( !m_queue.empty() ) { value = m_queue.front(); m_queue.pop(); rtn = true; } pthread_mutex_unlock( &m_mutex ); return rtn; } Unfortunately there are occasional issues that may be the fault of this code. That is, there are two threads and sometimes thread 1 never comes out of push() and at other times thread 2 never comes out of pop() (the block parameter is true) though the queue isn't empty. I understand there are other implementations available, but I'd like to try to fix this code, if needed. Anyone see any issues? The constructor has the appropriate initializations: Queue() { pthread_mutex_init( &mMutex, NULL ); pthread_cond_init( &mCondition, NULL ); } and the destructor, the corresponding 'destroy' calls.

  Read the article

• How do I conditionally compile C code snippets to my Perl module?

  - by mobrule

  I have a module that will target several different operating systems and configurations. Sometimes, some C code can make this module's task a little easier, so I have some C functions that I would like to bind the code. I don't have to bind the C functions -- I can't guarantee that the end-user even has a C compiler, for instance, and it's generally not a problem to failover gracefully to a pure Perl way of accomplishing the same thing -- but it would be nice if I could call the C functions from the Perl script. Still with me? Here's another tricky part. Just about all of the C code is system specific -- a function written for Windows won't compile on Linux and vice-versa, and the function that does a similar thing on Solaris will look totally different. #include <some/Windows/headerfile.h> int foo_for_Windows_c(int a,double b) { do_windows_stuff(); return 42; } #include <path/to/linux/headerfile.h> int foo_for_linux_c(int a,double b) { do_linux_stuff(7); return 42; } Furthermore, even for native code that targets the same system, it's possible that only some of them can be compiled on any particular configuration. #include <some/headerfile/that/might/not/even/exist.h> int bar_for_solaris_c(int a,double b) { call_solaris_library_that_might_be_installed_here(11); return 19; } But ideally we could still use the C functions that would compile with that configuration. So my questions are: how can I compile C functions conditionally (compile only the code that is appropriate for the current value of $^O)? how can I compile C functions individually (some functions might not compile, but we still want to use the ones that can)? can I do this at build-time (while the end-user is installing the module) or at run-time (with Inline::C, for example)? Which way is better? how would I tell which functions were successfully compiled and are available for use from Perl? All thoughts appreciated!

  Read the article

• multiple calls to realloc() seems to cause a heap corruption..

  - by Windindeed

  What's the problem with this code? It crashes every time. One time it's a failed assertion "_ASSERTE(_CrtIsValidHeapPointer(pUserData));", other times it is just a "heap corrpuption" error. Changing the buffer size affects this issue in some strange ways - sometimes it crashes on the "realloc", and other times on the "free". I have debugged this code many times, and there is nothing abnormal regarding the pointers. char buf[2000]; char *data = (char*)malloc(sizeof(buf)); unsigned int size = sizeof(buf); for (unsigned int i = 0; i < 5; ++i) { char *ptr = data + size; size += sizeof(buf); char *tmp = (char*)realloc(data, size); if (!tmp) { std::cout << "Oh no.."; break; } data = tmp; memcpy(ptr, buf, sizeof(buf)); } free(data); Thanks!

  Read the article

< Previous Page | 280 281 282 283 284 285 286 287 288 289 290 291  | Next Page >